1. Field
The present invention relates to a compact fluorescence detection instrument with optics for use in assays performed in a microfluidic cartridge.
2. Description of the Related Art
Although the benefits of the use of fluorophores as probes for in-vitro diagnostic assays are well known, the most commonly available forms of equipment for such assays are large, complex to use, relatively slow and rely on expensive confocal optics. These attributes make much equipment unsuitable for fully integrated “sample-to-answer” testing in remote locales and on-site at the point of care, where such equipment is required to be rugged, fast, compact, inexpensive, and easy to use. Although automated nucleic acid amplification in a microfluidic cartridge was first proposed some years ago (see Wilding, U.S. Pat. Nos. 5,304,487 and 5,635,358), detection of fluorescent assay targets outside controlled laboratory conditions is still hampered by the lack of portable and robust equipment. Two decades since their inception, molecular diagnostics are still relatively uncommon in the absence of advanced laboratory facilities because of these and other unsolved problems.
Needed to promote broader access to molecular diagnostics are self-contained assay systems designed to operate outside specialized laboratory facilities. Nucleic acid assays are rapidly becoming the “gold standard” for the detection of many different disease types, including infectious diseases, because they offer both higher sensitivity and specificity. Such assays have proven highly specific to a broad range of pathogenic conditions and are useful for tracking genetic strains of a particular disease as is fundamental to epidemiology, for example in discriminating H5N1 avian influenza from other types of influenza A or B, in determining whether a particular pathogen target is of a drug-resistant strain or not, and in detecting toxigenic strains of an enteric isolate such as E. coli O157:H7. Fluorescence-based assays have also been shown to be useful for monitoring conditions such as diabetes, cardiopathies, coagulopathies, immunoassays in general, and for detection of endotoxin in foods or drug products for example. Improved equipment is particularly needed for the large numbers of remote health clinics in the developing world where access to health care is limited and many infectious diseases are endemic, and health and life expectancy are poor.
In a typical fluorescence assay system, a fluorescent probe or fluorophore absorbs light having a wavelength or range of wavelengths and becomes excited; and the fluorophore then emits a fluorescent signal. The activity or inactivity of the fluorophore is indicative of the assay result. The emission signal has a wavelength or range of wavelengths that is generally longer than the exciting light (but may be shorter as in “up-converting fluorophores”). A dichroic beam splitter or band-pass filter, or combination thereof, is then used to separate the fluorescent signal from other light, and the signal is passed to a sensor. The sensor is often a photodiode, and generates an electrical signal that can be used to score the assay. Qualitative and quantitative assays using real time or endpoint fluorometry are feasible.
In such systems, a liquid sample is conveyed via a microfluidic channel into a detection chamber or channel of a microfluidic cartridge where a fluorescent probe admixed with or native to the sample is excited by an excitation source. Controls may be run in parallel or multiplexed in the assay channel. Emitted light is measured to determine the presence or absence of a target. A plurality of detection channels may be arranged in the detection region of the microfluidic cartridge. Assays involve making one or more measurements of fluorescence; fluorophores may be used as markers for nucleic acid amplicons formed in an amplification step, or more generally for the presence or absence of a fluorescent assay target. Real time fluorometry, FRET, qPCR, thermal melt curves, kinetic and rate endpoints for assay scoring and validation are also known in the art.
Prior art fluorescence detectors typically employ relatively expensive optical components (such as confocal optics, lasers and aspheric lenses) in order to pick up and localize fluorescent emissions present within a microfluidic cartridge or microarray. WO 98/049543 to Juncosa, for example, teaches three dichroic beam splitters in a single optical train, one for controlling excitation source power and another for controlling reflectance signal; the third dichroic beam splitter is used for discriminating probe-specific fluorescent emission. One or more lenses serve to focus the excitation beam on the sample. Juncosa further teaches use of an aperture at the inlet of a photomultiplier and optical objective lens components of a confocal microscope for controlling an imaging beam with a resolution of “microlocations” at about fifty microns. “By restricting the scope of the illumination to the area of a given microlocation, or a fraction thereof, coupled with restricting the field of view of the detector to the region of illumination, preferably through use of an aperture, significant improvements in signal-to-noise ratio may be achieved.” [p 7, lines 10-15]. These teachings are presaged by U.S. Pat. No. 3,013,467 to Minsky, U.S. Pat. Nos. 5,296,703 to Tsien, 5,192,980 to Dixon, 5,631,734 to Stern, 5,730,850 to Kambara, and are reiterated in U.S. Pat. No. 6,614,030 to Maher and U.S. Pat. No. 6,731,781 to Shams, among others. Maher uses lasers, fiber optics, a quartz plate and aspherical lenses with mini-confocal optical system in order to optimize focusing and emission at a ten micron-sized spot at the center of the microfluidic chamber [Col 3, lines 23-38; Col 7, lines 7-16, 43-48 and 58-63].
Similarly, in U.S. Pat. No. 6,635,487, Lee affirms that focusing the cone of the excitation beam on the plane of the sample “provides the greatest intensity to enhance analytical detection measurements on the assay chips.” [Col 1, lines 57-59]. This teaching thus encapsulates the prior art.
In a more recent filing, US Patent Application 2008/0297792 to Kim teaches that an image of an LED serving as a light source for fluorescence detection in a microfluidic chip is projected onto a sample as an “optical spot” by an objective lens. The optical spot is focused at the middle of the depth of a fluid in a chamber in the microfluidic chip [para 0018, 0067, claim 5]. Fluorescence emitted by the sample is collimated as nearly as possible to parallel rays by the objective lens and focused on an avalanche photodiode. The requirement for high precision in alignment relates to the dichroic mirror because the stopband will be shifted for light rays that do not enter the mirror at a 45° angle [para 0071], as is well known. Thus the teachings of Kim reflect the generally recognized state of the art.
In PCT Publication WO2008/101732 to Gruler, where is described a fluorescence detector head for multiplexing multiple excitation and detection wavelengths in a single light path, it is stated that, “A confocal measurement means that the focus of the illumination optics or the source, respectively intrinsically is the same as the focus of the detection optics or sensor, respectively.” [p7, lines 13-16]. Gruler goes on to state, “The confocal optics [of the invention] . . . secures highest signal and lowest background intrinsic features of confocal design” [p32, lines 1-5], i.e., according to Gruler the highest possible signal and lowest noise are obtained with confocal optics.
While the consensus teaching of the prior art arose out of the specialized use of confocal optics for epifluorescence microscopy, the teaching has been widely and uncritically applied to microfluidic, lateral flow, capillary electrophoresis and microarray applications. However, we have found that this approach is not well suited to liquid phase microfluidic diagnostic assays where detection of one or more molecular probes in a fluid-filled channel is required. Due to effects such as photoquenching, thermo-convection, and the occasional presence of bubbles or gradients in a fluid-filled channel, colocalizing the focal point of the excitation beam and emission cone in the plane of the sample chamber can lead to unacceptable instability, loss of signal, quenching, noise, irreproducibility and overall loss of sensitivity in the results. Because of the higher temperatures of PCR, for example, outgassing of reagents and sample is not an uncommon problem, and interference from bubbles entrained in the liquid sample is a frequent problem. The conventional approach also requires more expensive optical components and thus is disadvantageous for widespread application outside advanced clinical laboratories.
A second problem is assay validation. Current standards for validation of infectious disease assays by PCR, for example, have come to rely on use of spiked nucleic acid templates or more preferably, co-detection of endogenous normal flora, for example ubiquitous non-pathogenic Escherichia coli in stools where pathogens such as Salmonella typhi or E. coli O157 are suspected. Another ubiquitous endogenous template is human 18S rRNA, which is associated with higher quality respiratory and blood samples. Co-amplification and detection of an endogenous template ensures confidence in the assay results but is difficult to achieve in practice because of possible crosstalk between the fluorophores used as markers. When using high gain amplification, some level of crossover in the spectra of the excitation and emission of fluorophores commonly selected for multiplex PCR is typical and expected. Thus a solution that would isolate fluorescent signals with spectrally overlapping shoulders by using separate optical channels within a scanning detector head having shared low-noise electronics for downstream processing would be a technological advance of benefit in the art.
A third problem is portability. Use of disposable cartridges has proved beneficial because cross-contamination due to shared reagent reservoirs and shared fluid-contacting surfaces is avoided. However, configuring a precision optical instrument platform for accepting disposable cartridges is problematic. Problems include inaccuracies and stackup in mechanical tolerances that affect cartridge alignment and detector head positioning, the need for forming a highly conductive thermal interface between the plastic disposable cartridges and heating sources in the instrument, the need for sealing the pneumatic interface between control servos on the apparatus and microvalves on the cartridge, and the necessarily shorter light path available in a microfluidic cartridge (typically about or less than 1 millimeter), which without optimization can lead to loss in sensitivity. A simultaneous solution of these interlocking problems is only achieved by extensive experimentation and development, most often guided by trial and error in this highly unpredictable art. Thus there is a need in the art for numerous improvements, elements of which are the subject of the disclosure herein.